Optimizing Industrial Consumer Demand Response Through DIsaggregation, Hour-Ahead Pricing

141 Pages·2017·5.77 MB·English

by  Ahmed Abdulaal

University of Miami
Scholarly Repository
Open Access Dissertations Electronic Theses and Dissertations
2017-08-18
Optimizing Industrial Consumer Demand
Response Through DIsaggregation, Hour-Ahead
Pricing, and Momentary Autonomous Control
Ahmed Abdulaal
University of Miami, [email protected]
Follow this and additional works at:https://scholarlyrepository.miami.edu/oa_dissertations
Recommended Citation
Abdulaal, Ahmed, "Optimizing Industrial Consumer Demand Response Through DIsaggregation, Hour-Ahead Pricing, and
Momentary Autonomous Control" (2017).Open Access Dissertations. 1958.
https://scholarlyrepository.miami.edu/oa_dissertations/1958
This Open access is brought to you for free and open access by the Electronic Theses and Dissertations at Scholarly Repository. It has been accepted for
inclusion in Open Access Dissertations by an authorized administrator of Scholarly Repository. For more information, please contact
[email protected].
UNIVERSITY OF MIAMI

OPTIMIZING INDUSTRIAL CONSUMER DEMAND RESPONSE THROUGH
DISAGGREGATION, HOUR-AHEAD PRICING, AND MOMENTARY
AUTONOMOUS CONTROL

By
Ahmed Abdulaal

A DISSERTATION

Submitted to the Faculty
of the University of Miami
in partial fulfillment of the requirements for

the degree of Doctor of Philosophy

Coral Gables, Florida

August 2017
©2017
Ahmed Abdulaal
All Rights Reserved
UNIVERSITY OF MIAMI

A dissertation submitted in partial fulfillment of
the requirements for the degree of
Doctor of Philosophy

OPTIMIZING INDUSTRIAL CONSUMER DEMAND
RESPONSE THROUGH DISAGGREGATION, HOUR-AHEAD
PRICING, AND MOMENTARY AUTONOMOUS CONTROL

Ahmed Abdulaal

Approved:

________________ _________________
Shihab Asfour, Ph.D. Murat Erkoc, Ph.D.
Professor and Associate Dean Associate Professor of Industrial
College of Engineering Engineering

________________ _________________
Ramin Moghaddass, Ph.D. Moataz Eltoukhy, Ph.D.
Assistant Professor of Industrial Assistant Professor of
Engineering Kinesiology and Sport Sciences

________________ _________________
Osama Mohammed, Ph.D. Guillermo J. Prado, Ph.D.
Professor and Director of Energy Dean of the Graduate School
Systems Research Laboratory
Florida International University
ABDULAAL, AHMED (Ph.D., Industrial Engineering)
Optimizing Industrial Consumer Demand Response (August 2017)
Through Disaggregation, Hour-Ahead Pricing, and
Momentary Autonomous Control

Abstract of a dissertation at the University of Miami.

Dissertation supervised by Professor Shihab Asfour.
No. of pages in text. (129)

The work in this study addresses the current limitations of the price-driven demand
response (DR) approach. Mainly, the dependability on consumers to respond in an energy
aware conduct, the response timeliness, the difficulty of applying DR in a busy industrial
environment, and the problem of load synchronization are of utmost concern. In order to
conduct a simulation study, realistic price simulation model and consumers’ building load
models are created using real data. DR action is optimized using an autonomous control
method, which eliminates the dependency on frequent consumer engagement. Since load
scheduling and long-term planning approaches are infeasible in the industrial environment,
the proposed method utilizes instantaneous DR in response to hour-ahead price signals
(RTP-HA). Preliminary simulation results concluded savings at the consumer-side at the
cost of increased supplier-side burden due to the aggregate effect of the universal DR
policies. Therefore, a consumer disaggregation strategy is briefly discussed. Finally, a
refined discrete-continuous control system is presented, which utilizes multi-objective
Pareto optimization, evolutionary programming, utility functions, and bidirectional loads.
Demonstrated through a virtual testbed fit with real data, the new system achieves
momentary optimized DR in real-time while maximizing the consumer’s wellbeing.
ACKNOWLEDGEMENTS

I would like to thank my advisor, Dr. Shihab Asfour, for his encouragement,
guidance, financial, and emotional support. I would also like to thank Dr. Osama
Mohammed and Dr. Ramin Moghaddass for guiding my work and for their constructive
inputs.
I must also thank my coworker and dear friend, Jaime Buitrago, for his friendship,
technical assistance, and emotional support.
My appreciation also extends to the U.S. Department of Energy for funding the
University of Miami Industrial Assessment Center (MIIAC), which provided my doctoral
assistantship and made available all the equipment and data used in my studies.
My deepest thanks go to my family; my parents Reda Abdulaal and Fatma
Elbaghdady, my sister Sarah Abdulaal, and my wife Yuliya Krauchuk, for their
exceptional patience, love, and unconditional support. Without them, my achievements
would not have been possible.
Finally, I am very thankful to many of my friends, coworkers, students, and teachers
who I could not have mentioned all their names, but they have impacted my life during
the past few years with priceless knowledge and joyful experiences.

iii
TABLE OF CONTENTS

Page

LIST OF FIGURES ..................................................................................................... vi

LIST OF TABLES ....................................................................................................... ix

Chapter

1 INTRODUCTION ........................................................................................... 1
Demand-Supply Nature and Problems……………………………………… . 1
Motivation……………………………………………………………………. 2
Tariffs and RTP………………………………………………………………. 3
Research Objectives…………………………………………………………. 6
Research Scope and Reasoning………………………………………………. 6
Research Vision………………………………………………………………. 9

2 LITERATURE REVIEW ................................................................................ 10
RTP in Residential DR..................................................................................... 10
RTP in Industrial DR ....................................................................................... 12
Financial Implications and Price Setting in RTP ............................................. 13
Environmental Impact of RTP ......................................................................... 14
Load Synchronization ...................................................................................... 14
Remarks from Reviewed Literature ................................................................. 15

3 RTP TARIFF MODELING ............................................................................. 17
Chapter Introductory Remarks ......................................................................... 17
Chapter Objective and Motivation ................................................................... 20
Data Collection ................................................................................................ 20
Modeling Demand Forecast ............................................................................. 22
Modeling RTP Tariff ....................................................................................... 24

4 AUTONOMOUS LINEAR DEMAND CONTROL IN REAL-TIME .......... 29
Chapter Introductory Remarks ......................................................................... 29
Chapter Motivation .......................................................................................... 30
Chapter Objective ............................................................................................ 30
Operation Scheme ............................................................................................ 31
Mathematical Formulation ............................................................................... 33
Case Study: Industrial Facility Air Handling System ...................................... 38
Results and Discussion .................................................................................... 44
Chapter Conclusive Remarks ........................................................................... 51

5 LOAD DISAGGREGATION .......................................................................... 53
Chapter Introductory Remarks ......................................................................... 53
iv
Chapter Motivation .......................................................................................... 53
Chapter Objectives ........................................................................................... 54
Data for Disaggregation ................................................................................... 54
Clustering Load Profiles .................................................................................. 55
Classification of DR Model Data ..................................................................... 58
The Fuzzy Genetic Algorithm for Classification ............................................. 59
Conclusions and Remarks for Application to DR ............................................ 61

6 CONTINUOUS MULTI-OBJECTIVE DEMAND RESPONSE TO DISCRETE
CONTROLLER SIGNALS ............................................................................. 64
Chapter Introductory Remarks ......................................................................... 64
Chapter Motivation and Objectives ................................................................. 67
Stage 1: Load Shifting Targets Optimization in Discrete Time ...................... 67
Utility Functions for Controllable Loads ......................................................... 71
Stage 2: Dynamic Multi-Objective Load Management in Real-Time ............. 79
Evolutionary Programming for Real-Time Nonlinear Optimization ............... 84
Pareto Frontier Analysis .................................................................................. 87

7 SIMULATION MODELING AND RESULTS .............................................. 90
Chapter Introductory Remarks ......................................................................... 90
Chapter Motivation and Objective ................................................................... 90
Real Environment Data Collection .................................................................. 91
Model Parameter Estimation............................................................................ 92
Model Communication Architecture ............................................................... 94
AHP for Weights Assignments ........................................................................ 99
Simulation Model Input Data........................................................................... 100
Results and Analysis ........................................................................................ 102

8 CONCLUSIONS.............................................................................................. 112
Summary .......................................................................................................... 112
Contributions.................................................................................................... 115
Limitations ....................................................................................................... 117
Implementation and Future Work .................................................................... 118

REFERENCES…………… ........................................................................................ 121
v
LIST OF FIGURES

Page

Figure 1.1. Electricity demand fluctuations. (a) by month. (b) by day. (c) by hour.
Sampled data downloaded from PJM RTO for the year 2015 [2]. .............................. 2

Figure 1.2. Estimated energy flow in the US for the year 2015. Source: Lawrence
Livermore National Laboratory [49]. .......................................................................... 8

Figure 3.1. System load in kW used in the simulation model. (a) 1-minute load per
consumer. (b) 1-minute total system load. (c) Hourly average load per consumer. (d)
hourly average total system load. ................................................................................. 22

Figure 3.2. SIMULINK block diagram for generating forecasted demand from actual
demand data in day-ahead forecasting. ........................................................................ 23

Figure 3.3. Actual demand and simulated forecasted demands. .................................. 23

Figure 3.4. Path diagram for price prediction model. .................................................. 27

Figure 3.5. Load, actual, and simulated spot market prices (Mid-Atlantic region data for
two weeks starting 7/14/2014). .................................................................................... 28

Figure 4.1. Flow chart of demand shifting controller in a RTP-HA scheme where price
information are available for the current hour and the following hour only. .............. 32

Figure 4.2. Building MI0189 1-week demand profile with 5 candidate HVAC
components for demand control: (a) 1-minute logged demand and (b) hourly averaged
demand. .............................................................................................................. 39

Figure 4.3. The Proposed controller and regulator system framework in SIMULINK for
an industrial building with EMS and controllable cooling/ventilation load. ............... 40

Figure 4.4. Regulator SIMULINK block schema for adjusting thermostat setpoints based
on controller signal (e). ................................................................................................ 41

Figure 4.5. Regulator SIMULINK block schema for cycling between parallel HVAC
units based controller signal (e). .................................................................................. 42

Figure 4.6. Duct system demand profiles before implementing the proposed controller
(top) and after (bottom). ............................................................................................... 43

Figure 4.7. Comparison of the indoor climate conditions before and after controller
implementation. ........................................................................................................... 44
vi
Figure 4.8. The eight consumers’ demand Profiles prior and Post to demand shifting in
the RTP-HA scheme. ................................................................................................... 46

Figure 4.9. The change in the lumped equipment load profiles for the eight due to load
shifting in RTP-HA scheme. ........................................................................................ 47

Figure 4.10. New load synchronization peaks when all consumers use demand shifting
controllers at the exact same time instances following the RTP-HA signals. ............. 49

Figure 4.11. Reduction in load synchronization peaks when consumers use demand
shifting controllers at randomly offset time instances. ................................................ 49

Figure 5.1. Data clusters and cluster centroids using the wave function approach. .... 56

Figure 5.2. Clustered 1-month data plotted against their representative cluster wave
function. .............................................................................................................. 57

Figure 5.3. Clustered 1-week data plotted against their representative cluster centroid
using the time-series approach. .................................................................................... 58

Figure 5.4. Schematic representation of the FGA evolution processes. ...................... 61

Figure 5.5. Less variability and fewer load synchronization peaks when all consumers use
demand shifting controllers at the exact same time instances following the cluster-
disaggregated RTP-HA signals. ................................................................................... 63

Figure 6.1. Pseudo-code for the hierarchical optimization scheme (showing only stage-1
in detail), where H is the planning horizon (e.g. H=24 hours). ................................... 71

Figure 6.2. Multistage cooling unit utility function charts for various settings. .......... 74

Figure 6.3. EV charging utility function charts for various EV attributes. ................. 76

Figure 6.4. EV V2B utility function charts for various price preferences and assuming all
other attributes are held constant. ................................................................................ 79

Figure 6.5. Chromosome coding for customized binary GA. ...................................... 85

Figure 6.6. Chromosomes combination during crossover. .......................................... 86

Figure 6.7. Chromosome alteration during mutation. .................................................. 86

Figure 7.1. Logged cooling loads vs. fitted 4-stage chiller system. ............................. 92

Figure 7.2. SIMULINK thermal parameters estimation model. .................................. 93
vii